/*
* Found a sequence mapping. Locate the start/end mapped residues.
*/
- List<AlignedCodonFrame> mapping = Arrays.asList(new AlignedCodonFrame[] { acf });
+ List<AlignedCodonFrame> mapping = Arrays
+ .asList(new AlignedCodonFrame[] { acf });
SearchResults sr = buildSearchResults(selected,
startResiduePos, mapping);
for (Match m : sr.getResults())
char fromGapChar = mapFrom.getAlignment().getGapCharacter();
- // FIXME allow for hidden columns
-
/*
* For each mapped column, find the range of columns that residues in that
* column map to.
*/
- for (Object obj : colsel.getSelected())
+ List<SequenceI> fromSequences = mapFrom.getAlignment().getSequences();
+ List<SequenceI> toSequences = mapTo.getAlignment().getSequences();
+
+ for (Integer sel : colsel.getSelected())
+ {
+ mapColumn(sel.intValue(), codonFrames, mappedColumns, fromSequences,
+ toSequences, fromGapChar);
+ }
+
+ for (int[] hidden : colsel.getHiddenColumns())
+ {
+ mapHiddenColumns(hidden, codonFrames, mappedColumns, fromSequences,
+ toSequences, fromGapChar);
+ }
+ return mappedColumns;
+ }
+
+ /**
+ * Helper method that maps a [start, end] hidden column range to its mapped
+ * equivalent
+ *
+ * @param hidden
+ * @param mappings
+ * @param mappedColumns
+ * @param fromSequences
+ * @param toSequences
+ * @param fromGapChar
+ */
+ protected static void mapHiddenColumns(int[] hidden,
+ List<AlignedCodonFrame> mappings, ColumnSelection mappedColumns,
+ List<SequenceI> fromSequences, List<SequenceI> toSequences,
+ char fromGapChar)
+ {
+ for (int col = hidden[0]; col <= hidden[1]; col++)
{
- int col = ((Integer) obj).intValue();
- int mappedToMin = Integer.MAX_VALUE;
- int mappedToMax = Integer.MIN_VALUE;
+ int[] mappedTo = findMappedColumns(col, mappings, fromSequences,
+ toSequences, fromGapChar);
/*
- * For each sequence in the 'from' alignment
+ * Add the range of hidden columns to the mapped selection (converting
+ * base 1 to base 0).
*/
- for (SequenceI fromSeq : mapFrom.getAlignment().getSequences())
+ if (mappedTo != null)
{
- /*
- * Ignore gaps (unmapped anyway)
- */
- if (fromSeq.getCharAt(col) == fromGapChar)
- {
- continue;
- }
+ mappedColumns.hideColumns(mappedTo[0] - 1, mappedTo[1] - 1);
+ }
+ }
+ }
+
+ /**
+ * Helper method to map one column selection
+ *
+ * @param col
+ * the column number (base 0)
+ * @param mappings
+ * the sequence mappings
+ * @param mappedColumns
+ * the mapped column selections to add to
+ * @param fromSequences
+ * @param toSequences
+ * @param fromGapChar
+ */
+ protected static void mapColumn(int col,
+ List<AlignedCodonFrame> mappings, ColumnSelection mappedColumns,
+ List<SequenceI> fromSequences, List<SequenceI> toSequences,
+ char fromGapChar)
+ {
+ int[] mappedTo = findMappedColumns(col, mappings, fromSequences,
+ toSequences, fromGapChar);
+
+ /*
+ * Add the range of mapped columns to the mapped selection (converting
+ * base 1 to base 0). Note that this may include intron-only regions which
+ * lie between the start and end ranges of the selection.
+ */
+ if (mappedTo != null)
+ {
+ for (int i = mappedTo[0]; i <= mappedTo[1]; i++)
+ {
+ mappedColumns.addElement(i - 1);
+ }
+ }
+ }
+
+ /**
+ * Helper method to find the range of columns mapped to from one column.
+ * Returns the maximal range of columns mapped to from all sequences in the
+ * source column, or null if no mappings were found.
+ *
+ * @param col
+ * @param mappings
+ * @param fromSequences
+ * @param toSequences
+ * @param fromGapChar
+ * @return
+ */
+ protected static int[] findMappedColumns(int col,
+ List<AlignedCodonFrame> mappings, List<SequenceI> fromSequences,
+ List<SequenceI> toSequences, char fromGapChar)
+ {
+ int[] mappedTo = new int[] { Integer.MAX_VALUE, Integer.MIN_VALUE };
+ boolean found = false;
+
+ /*
+ * For each sequence in the 'from' alignment
+ */
+ for (SequenceI fromSeq : fromSequences)
+ {
+ /*
+ * Ignore gaps (unmapped anyway)
+ */
+ if (fromSeq.getCharAt(col) == fromGapChar)
+ {
+ continue;
+ }
+
+ /*
+ * Get the residue position and find the mapped position.
+ */
+ int residuePos = fromSeq.findPosition(col);
+ SearchResults sr = buildSearchResults(fromSeq, residuePos, mappings);
+ for (Match m : sr.getResults())
+ {
+ int mappedStartResidue = m.getStart();
+ int mappedEndResidue = m.getEnd();
+ SequenceI mappedSeq = m.getSequence();
/*
- * Get the residue position and find the mapped position.
+ * Locate the aligned sequence whose dataset is mappedSeq. TODO a
+ * datamodel that can do this efficiently.
*/
- int residuePos = fromSeq.findPosition(col);
- SearchResults sr = buildSearchResults(fromSeq, residuePos,
- codonFrames);
- for (Match m : sr.getResults())
+ for (SequenceI toSeq : toSequences)
{
- int mappedStartResidue = m.getStart();
- int mappedEndResidue = m.getEnd();
- SequenceI mappedSeq = m.getSequence();
-
- /*
- * Locate the aligned sequence whose dataset is mappedSeq. TODO a
- * datamodel that can do this efficiently.
- */
- for (SequenceI toSeq : mapTo.getAlignment().getSequences())
+ if (toSeq.getDatasetSequence() == mappedSeq)
{
- if (toSeq.getDatasetSequence() == mappedSeq)
- {
- int mappedStartCol = toSeq.findIndex(mappedStartResidue);
- int mappedEndCol = toSeq.findIndex(mappedEndResidue);
- mappedToMin = Math.min(mappedToMin, mappedStartCol);
- mappedToMax = Math.max(mappedToMax, mappedEndCol);
- // System.out.println(fromSeq.getName() + " mapped to cols "
- // + mappedStartCol + ":" + mappedEndCol);
- break;
- // note: remove break if we ever want to map one to many sequences
- }
+ int mappedStartCol = toSeq.findIndex(mappedStartResidue);
+ int mappedEndCol = toSeq.findIndex(mappedEndResidue);
+ mappedTo[0] = Math.min(mappedTo[0], mappedStartCol);
+ mappedTo[1] = Math.max(mappedTo[1], mappedEndCol);
+ found = true;
+ break;
+ // note: remove break if we ever want to map one to many sequences
}
}
}
- /*
- * Add the range of mapped columns to the mapped selection (converting
- * base 1 to base 0). Note that this may include intron-only regions which
- * lie between the start and end ranges of the selection.
- */
- for (int i = mappedToMin; i <= mappedToMax; i++)
- {
- mappedColumns.addElement(i - 1);
- }
}
- return mappedColumns;
+ return found ? mappedTo : null;
}
/**
public static List<AlignedCodonFrame> findMappingsForSequence(
SequenceI sequence, List<AlignedCodonFrame> mappings)
{
+ return findMappingsForSequenceAndOthers(sequence, mappings, null);
+ }
+
+ /**
+ * Returns a list of any mappings that are from or to the given (aligned or
+ * dataset) sequence, optionally limited to mappings involving one of a given
+ * list of sequences.
+ *
+ * @param sequence
+ * @param mappings
+ * @param filterList
+ * @return
+ */
+ public static List<AlignedCodonFrame> findMappingsForSequenceAndOthers(
+ SequenceI sequence, List<AlignedCodonFrame> mappings,
+ List<SequenceI> filterList)
+ {
List<AlignedCodonFrame> result = new ArrayList<AlignedCodonFrame>();
if (sequence == null || mappings == null)
{
{
if (mapping.involvesSequence(sequence))
{
- result.add(mapping);
+ if (filterList != null)
+ {
+ for (SequenceI otherseq : filterList)
+ {
+ SequenceI otherDataset = otherseq.getDatasetSequence();
+ if (otherseq == sequence
+ || otherseq == sequence.getDatasetSequence()
+ || (otherDataset != null && (otherDataset == sequence || otherDataset == sequence
+ .getDatasetSequence())))
+ {
+ // skip sequences in subset which directly relate to sequence
+ continue;
+ }
+ if (mapping.involvesSequence(otherseq))
+ {
+ // selected a mapping contained in subselect alignment
+ result.add(mapping);
+ break;
+ }
+ }
+ }
+ else
+ {
+ result.add(mapping);
+ }
}
}
return result;
}
+
+ /**
+ * Returns the total length of the supplied ranges, which may be as single
+ * [start, end] or multiple [start, end, start, end ...]
+ *
+ * @param ranges
+ * @return
+ */
+ public static int getLength(List<int[]> ranges)
+ {
+ if (ranges == null)
+ {
+ return 0;
+ }
+ int length = 0;
+ for (int[] range : ranges)
+ {
+ if (range.length % 2 != 0)
+ {
+ System.err.println("Error unbalance start/end ranges: "
+ + ranges.toString());
+ return 0;
+ }
+ for (int i = 0; i < range.length - 1; i += 2)
+ {
+ length += Math.abs(range[i + 1] - range[i]) + 1;
+ }
+ }
+ return length;
+ }
+
+ /**
+ * Answers true if any range includes the given value
+ *
+ * @param ranges
+ * @param value
+ * @return
+ */
+ public static boolean contains(List<int[]> ranges, int value)
+ {
+ if (ranges == null)
+ {
+ return false;
+ }
+ for (int[] range : ranges)
+ {
+ if (range[1] >= range[0] && value >= range[0] && value <= range[1])
+ {
+ /*
+ * value within ascending range
+ */
+ return true;
+ }
+ if (range[1] < range[0] && value <= range[0] && value >= range[1])
+ {
+ /*
+ * value within descending range
+ */
+ return true;
+ }
+ }
+ return false;
+ }
+
+ /**
+ * Removes a specified number of positions from the start of a ranges list.
+ * For example, could be used to adjust cds ranges to allow for an incomplete
+ * start codon. Subranges are removed completely, or their start positions
+ * adjusted, until the required number of positions has been removed from the
+ * range. Reverse strand ranges are supported. The input array is not
+ * modified.
+ *
+ * @param removeCount
+ * @param ranges
+ * an array of [start, end, start, end...] positions
+ * @return a new array with the first removeCount positions removed
+ */
+ public static int[] removeStartPositions(int removeCount,
+ final int[] ranges)
+ {
+ if (removeCount <= 0)
+ {
+ return ranges;
+ }
+
+ int[] copy = Arrays.copyOf(ranges, ranges.length);
+ int sxpos = -1;
+ int cdspos = 0;
+ for (int x = 0; x < copy.length && sxpos == -1; x += 2)
+ {
+ cdspos += Math.abs(copy[x + 1] - copy[x]) + 1;
+ if (removeCount < cdspos)
+ {
+ /*
+ * we have removed enough, time to finish
+ */
+ sxpos = x;
+
+ /*
+ * increment start of first exon, or decrement if reverse strand
+ */
+ if (copy[x] <= copy[x + 1])
+ {
+ copy[x] = copy[x + 1] - cdspos + removeCount + 1;
+ }
+ else
+ {
+ copy[x] = copy[x + 1] + cdspos - removeCount - 1;
+ }
+ break;
+ }
+ }
+
+ if (sxpos > 0)
+ {
+ /*
+ * we dropped at least one entire sub-range - compact the array
+ */
+ int[] nxon = new int[copy.length - sxpos];
+ System.arraycopy(copy, sxpos, nxon, 0, copy.length - sxpos);
+ return nxon;
+ }
+ return copy;
+ }
}